Autor(es): Costa, M. ; Rocha, T. ; Mendonça, J. ; Pilão, R. ; Pinto, P.
Data: 2024
Identificador Persistente: http://hdl.handle.net/10400.22/29534
Origem: Repositório Científico do Instituto Politécnico do Porto
Assunto(s): Wind energy; post-construction Energy Yield Assessment; linear regression
The uncertainty associated with the prospective Energy Yield Assessment (EYA) of a wind farm may be reduced by re estimating the energy yield after it enters normal operation. This study aims to validate a simple methodology for conducting post-construction EYA of an operational wind farm. The proposed methodology derives a linear relationship between a historical source of wind speed data and the observed wind farm production on a monthly basis. In a first stage, the impact of different data sources on the accuracy of the Long-Term energy yield estimate was assessed. Results suggest that the determination coefficient R 2 is a reliable indicator for selecting the most adequate source of historical wind speed data to be used in the Long-Term energy yield estimate. In a second stage, the model was validated from a statistical point of view by testing the premises of the linear regression model, namely the significance of the linear correlation (ANOVA test), and normally-distributed (Shapiro-Wilk test), non-self-correlated (Durbin-Watson), homoscedastic (Breusch-Pagan test) residuals. Results show these premises are verified for most test cases, indicating that the model is statistically robust that the model is statistically robust for most test cases.
